虹膜斑块特征提取方法研究

Study on Extraction Method of Plaque Characteristics

虹膜斑块的形状和位置可以反映人体器官的健康状况。针对虹膜诊病中斑块的自动提取问题,采用一种计算机视觉技术自动检测虹膜斑块的方法。利用预处理的相关方法提取虹膜部分,通过一个窗口找到斑块所在区域,在斑块区域内的中心像素灰度值的平均值小于边缘像素灰度值的最小值,用水平集方法,得到这些替代区域的边缘信息,观察零水平集轮廓是否满足封闭条件和标准尺寸准则。如果满足,则认定是斑块的边缘。通过多次实验检验,证明了该方法适用于色素斑背景不均匀的虹膜图像,具有很好的效果。

The shape and location of iris plaques can reflect the health status of human organs. In view of automatic extraction for medical consultation of iris plaques in question,propose a method of automatic detection of iris plaque using computer vision techniques. Use re-lated pretreatment method to extract the iris section,through a window to find the plaque area. The average pixel value in the patch area of gray value is less than the minimum value of the pixel gray edge. Using level set methods,these alternative area edge information can be get. Observe the zero level set contour whether meets the closure condition and standard size criterion or not. If satisfied,that is patch edge. Through several experiments testing,show that this method is suitable for complex background iris image,with good effect.